On Sun, 5 Sep 2004, Tim Jenness wrote:
> - the ams test fails. But then it fails on my 32bit system. Does the ams
> test in pcs make check work for anyone?
No, it's never worked for me. The amsmastertest script should work
standalone (the error is because amsslave and amsmaster commit suicide and
those signals are being passed on, don't know why they do that rather than
just exit).
> I have to say that I'm appalled at the amount of work it is taking to port
> Starlink to 64bit pointers given that it's been running on an alpha for
> years. Those alpha compiler options to fit all pointers into 32bit made
> life too easy! Only a handful of libraries had been switched.
But imagine how much work it would be without that background.
> As for applications, I've done atools, sx, datacube, convert and kappa.
> The AMD system has successfully displayed an image and calculated stats. I
> haven't tried the demo yet since remote display onto my home system is
> painful.
>
> Any chance that some of the slack can be taken up by others here for the
> remaining apps or do I have to do them all myself?
I've started CNFing the GAIA C/Fortran interface pointer handling, but
progress is slow as the number of pointers heading into Fortran is larger
than I anticipated, and, unlike other apps, they don't all come from HDS
or CNF.
Cheers,
Peter.
|